Data economy: A discussion on value,fictitious valorisation,and national sovereignty

Abstract

In this article, we present elements for a critique of the notion of a new data economy, often portrayed as a distinct stage in capitalist development, allegedly detached from the model of flexible accumulation characteristic of the finance-dominated phase of regulation. From a Marxist perspective, we argue that data should be understood as crude raw material, devoid of value. It is only when mobilised by labour that they are transformed into informational raw material, capable of being incorporated into different production processes – as in the case of the audience-commodity – or of presenting themselves as a commodity in their own right: the information-commodity. We also put forward the hypothesis that, when simply extracted and traded as promises of future use in productive processes, crude data configure a new form of fictitious accumulation. Finally, we reflect on the oligopolistic control exercised by the big techs, most of them based in the United States, over immense volumes of data, as well as on the implications this dominance bears for the sovereignty of peripheral countries.

Introduction

The process of productive restructuring and the consolidation of neoliberalism constitute responses by monopoly capital to the structural crisis that erupted in the 1970s. Richard Nixon's 1971 decision to end the convertibility of the dollar into gold brought the post-war economic order to a close and exposed the fragility of the international system based on the Bretton Woods agreements. This was, in fact, a single crisis articulated along two interrelated dimensions: the collapse of the Fordist-Keynesian accumulation pattern and the crisis of United States hegemony. This dual shock decisively undermined the mode of regulation established in the aftermath of the Second World War (Aglietta, 1979; Boyer, 1986).

Maria da Conceição Tavares (1985) was a pioneer in analysing the strategy for the restoration of US hegemony, identifying the macroeconomic, political and military mechanisms deployed by Ronald Reagan following the deflationary adjustment carried out by Paul Volcker under the previous administration. She characterised this strategy as a ‘bastard Keynesianism’ – or, in Perry Anderson's (1995) terms, a ‘disguised military Keynesianism’. However, the effort to reassert US hegemony also unfolded on the terrain of industrial policy, particularly during the Clinton administration, with the Global Information Infrastructure project (Bolaño, 2012), part of the wave of the micro-electronic revolution (Coutinho, 1992), which propelled the spread of information and communication technologies and consolidated the digital paradigm across the globe.

This transformation is essentially marked, on the one hand, by a broad process of subsumption and relative deskilling of intellectual labour ¹ and, on the other hand, by the general intellectualisation of all labour processes, generating profound impacts on consumption, culture and ways of life (Bolaño, 2002). In Marxist terms, this process can be defined as the Third Industrial Revolution, in which software assumes the role formerly played by the machine-tool during the original Industrial Revolution (Bolaño, 2002). Telematic convergence, the internet, digital platforms, so-called artificial intelligence, and other sociotechnical innovations of the past five decades are all developments of this same revolutionary process.

Our general objective is to present key elements for a critique of the idea of a new ‘data economy’, often portrayed as a new stage of capitalist development, seemingly detached from the model of flexible accumulation (Harvey, 1992), characteristic of the finance-dominated phase of regulation (Chesnais, 1996). We begin from the hypothesis that data have no intrinsic value. Only when mobilised by labour do they become informational raw material, capable of being incorporated into different production processes – as in the case of the audience-commodity – or of presenting themselves as a commodity in their own right: the information-commodity. We further argue that, when simply extracted and traded as promises of use in future productive processes, crude data configure a new modality of fictitious accumulation. Finally, we contend that the power of big techs over data, combined with the technological dependence of peripheral countries such as Brazil, heightens the risks to national sovereignty.

This article is organised into four sections, in addition to this introduction and the concluding remarks. The first addresses the issue of collecting and organising data that make up large digital repositories, which subsequently become inputs for different processes – whether the production of knowledge, commodities, organisation or surveillance. Unlike much of the literature on the subject, which conceives data as raw material (Srnicek, 2017) or as a commodity (Fuchs, 2014), we argue that data should instead be understood, from a Marxian perspective, as raw matter without any value until mobilised by living labour – in this case, informational labour. Building on the argument developed in the first section, the second offers a critique of the hypothesis of free digital labour (Dantas and Raulino, 2020; Fuchs, 2014). The third section shows that data, in addition to playing a central role in various concrete productive processes, may also be traded speculatively, thereby constituting a modality of fictitious capital accumulation. Finally, the fourth section discusses the power of big tech companies and their implications for national sovereignty.

Data: Neither raw material nor commodity

On one hand, the so-called data economy is a technological development linked to the Third Industrial Revolution, expanding the capacities for collecting, storing and manipulating enormous volumes of data. At this point, for methodological reasons, we are interested in focusing strictly on the economic aspects of the problem, but we cannot overlook the fact that among the destinations of the data extracted from populations are, prominently, surveillance and social control systems operated by agents endowed with ‘economic power’ (Furtado, 1978). In fact, this is the ultimate meaning of the entire process of constituting so-called big data: control.

In the manipulation of large numbers, algorithmic models play a crucial role in facilitating control and communication. Communication is understood here in the precise Marxist sense of the term (Bolaño, 2000), which becomes more complex when considering the specificities of the new structure of social mediation established with the emergence of digital platforms, human–machine relationships, and the automation of data collection and transfer by machines, independent of direct human action. A particularly relevant aspect, especially regarding the control of industrial labour processes, is the unit of measurement between man and machine (Sohn-Rethel, 1989), which, in effect, is at the origin of digital systems that represent the dialectical overcoming of the old Taylorism-Fordism. The expansion of the digitisation paradigm to other areas and other control mechanisms derives from this process.

On the other hand, as interest in statistical, demographic and behavioural data increases – for advertising strategies, political propaganda and much more, including the well-known training of artificial intelligences – the notion of a ‘data economy’ may appear increasingly fitting from the perspective of orthodox economics, which abstracts from social relations and makes no distinction between price and value, relying essentially on the subjective notion of utility.

An example of this orthodox perspective can be found in Marielza Oliveira (2023: 1), who states that ‘data […] is considered relational by definition and produces greater value when insights can be obtained about the entire community it represents’. Such an interpretation, however, is not grounded in the social relations between human beings embedded in an economy structured by the generalised production of commodities. From this perspective, the value of data does not result from the labour embodied in them, but rather emerges from their capacity to be cross-referenced, aggregated and analysed as a whole. This capacity, however, is not intrinsic to the data – it results from the labour that organises and processes them. What is attributed to data, therefore, is a utility-value: the greater their capacity to generate practical utilities, the greater their supposed value. In essence, this is a utilitarian conception, which abstracts from the relations of exploitation between labour and capital in the realm of digital platforms.

In opposition to these fetishised views, which attribute an intrinsic and natural value to things and ignore the historically specific dimension of labour and value within capitalist sociability, we argue that data should be defined neither as a commodity nor as raw material – as has become naturalised in the literature. Srnicek (2017) himself, when criticising authors who classify data as a commodity, rightly observes that ‘just like oil, data are a material to be extracted, refined, and used in a variety of ways’ (p. 40). However, even if his intention is correct, the author ends up characterising them – perhaps somewhat unreflectively – as raw material, instead of emphasising their condition as crude raw material without value.

Indeed, if oil only becomes raw material once it has been extracted and refined, data too only assume this condition after a similar process of extraction and refinement, as the author himself suggests. While they remain merely collected and stored in large repositories, data are nothing more than crude raw material, with no added value whatsoever.

To clarify this point, it is useful to return to Marx's concept of stock labour. In volume II of Capital, Marx distinguishes three modalities: ‘in the form of productive capital, of the individual fund of consumption, and of the stock of commodities or commodity-capital’ (Marx, 2014: 220). As Cotrim (2012: 69–70) explains, ‘only the first two forms of stock employ productive labour, since, being necessary to the process of producing use-values which they preserve, they add labour time to them’. Thus, whereas the stock of commodities strictly speaking is linked only to circulation – that is, to the transfer of property rights and the associated costs – in the first two forms what becomes evident is the action of labour upon the use-value, ensuring its preservation and preventing the loss of value already produced.

In the case of the storage of data as crude raw material, the situation is different: since data in themselves have no value, one cannot properly speak of productive labour devoted to preserving something that does not yet exist. What occurs is merely the conservation of a vast mass of unstructured data. However, once they undergo processes of transformation – being organised, structured and converted into inputs for other activities – they acquire the condition of informational raw material. It is at this point that stock labour, understood in the Marxian sense, can be associated with productive labour, insofar as it acts both to preserve the structuring already undertaken, preventing data from regressing to the condition of dispersed traces, and to expand it.

In this context of data as crude raw material, we can draw an analogy with what Marx (1959: 548) calls the ‘free gift of Nature to capital’. Just as natural resources that cost capital nothing, so too the digital traces produced by users and appropriated by platforms entail no direct cost. According to Marx (1959):

Natural elements entering as agents into production, and which cost nothing, no matter what role they play in production, do not enter as components of capital, but as a free gift of Nature to capital, that is, as a free gift of Nature's productive power to labour, which, however, appears as the productiveness of capital, as all other productivity under the capitalist mode of production.

In the analysed context, it is observed that data collection from internet users has been increasing more and more due to the ‘expanding platform infrastructures in the form of apps, plugins, active and passive sensors, and trackers’ (Poell et al., 2019: 6). However, these digital infrastructures have the sole purpose of capturing data, that is, separating them from their immediate connection with individuals. According to Marx (1996: 188), ‘all those things which labour merely separates from immediate connection with their environment, are subjects of labour spontaneously provided by Nature’. The same reasoning can be applied to data: they only become raw material endowed with value once they have ‘undergone some alteration by means of labour’ (Marx, 1996), that is, after being filtered, organised and structured by human labour, in stages subsequent to mere capture (Zanghelini, 2024).

Digital data, as they constitute crude raw material, can be compared to the mapping of the human genome, which produces a vast database of biological and genetic information of great utility. Bolaño (2003) emphasises this idea in his study on the genome project, mentioning both databases and clone libraries preserved for future experiments. In both cases, the transformation into raw material, that is, the initial valorisation, depends on retrieving the data from the repositories through the teleological action that characterises human labour in general (Lukács, 2013). In the case of the informational worker, the action is oriented by a project in which the industrial cycle and the cycle of certified academic production are often articulated (Bolaño, 2003). However, it should be noted that we are no longer dealing with data but, considering organised knowledge, with information, which begins to circulate within collective processes of physical and intellectual labour (Bolaño, 2000).

These repositories fulfil a dual function: they serve both subsequent genomic research and capitalist accumulation, through specific productive processes that result in new drugs, treatments and procedures with economic value. The mapping can be interpreted as a form of enclosure (Bolaño, 2003). In the historical case of land enclosure, it was a violent action, mediated by specific circumstances, that converted communal property into private property without adding any value. The subsequent use of this land in capitalist productive processes, through the labour of wage workers in agriculture, is a distinct matter. In the case of the Human Genome Project, a consensus was reached that genes themselves could not be patented, only the products developed through subsequent research. It is from this perspective that we consider a similar treatment should be applied to the databases controlled by the companies that own digital platforms.

A critique of the hypothesis of free digital labour

Based on what was discussed in the previous section, it is relevant to address Christian Fuchs’ (2014) problematic interpretation. According to the author, who bases his analysis on the controversial notion of audience labour by Dallas Smythe (1983), internet users themselves would be part of the collective labour responsible for value production, whose measure remains socially necessary labour time. Thus, internet users – including capitalist individuals – would be involved in producing the ‘data commodity’, later marketed by digital platforms to advertisers, even when performing completely mundane activities on the web.

Based on this extrapolation, the author further argues that prosumers – a neologism created by futurist Alvin Toffler (1980) to describe the fusion between producer and consumer activities – are being exploited, even without any relation of buying and selling labour power. For Fuchs, this means that the degree of labour exploitation in contemporary capitalism would be tending towards infinity, since online activities would be considered unpaid labour. Now, if this conception of unpaid digital labour, which finds a significant audience in academia, made any sense from the perspective of Marxian value theory, capitalism would have long since overcome its structural crisis; perhaps we would be in a new ‘golden phase’. In Fuchs’ (2014: 104) own words:

The rate of exploitation (also called the rate of surplus value) measures the relationship of workers’ unpaid work time and paid work time. The higher the rate of exploitation, the more work time is unpaid. Users of commercial social media platforms have no wages (v = 0). None of their usage time is remunerated in order to fund subsistence. Therefore the rate of surplus value converges towards infinity. Internet prosumer labour is infinitely exploited by capital. This means that capitalist prosumption is an extreme form of exploitation in which the prosumers work completely for free.

Dantas and Raulino (2020), albeit through a different route, make an error similar to that of Fuchs. They argue that the profits of advertising platforms such as Facebook or YouTube, classified as ‘informational monopolies’, would derive from an ‘informational rent’ generated by productive labour carried out both by waged workers and by the mass of unpaid internet users, without, however, a commodity being produced – unlike in Fuchs' account. More precisely, they claim that ‘unpaid labour provides a wide range of data to be valorised by capital’, while ‘waged labour, in turn, by controlling systems, algorithms and social research, actively seeks to keep free labour intensively mobilised’ (Dantas and Raulino, 2020: 131). In the authors’ own words:

We argue that socio-digital platforms (SDPs) do not produce commodities, even though the exploitation of the socially combined informational labour of both contracted professionals and their vast audience of millions or billions of users generates value and surplus value, which becomes the activity itself, or living labour. The surplus value is appropriated through the algorithmic “enclosure” of data, whether through the legal system of intellectual property or through a business model known in corporate jargon as “walled gardens” (Dantas and Raulino, 2020: 126).

This approach revisits the controversial interpretation of Herscovici (2014) concerning rent derived from intellectual property, according to which such rent would not be unproductive. ² In a rather unilluminating way, Dantas and Raulino (2020) introduce the notion of a ‘loan of data’ to advertisers and, drawing on section IV of volume III of Capital, claim that the gain in the platform economy would be analogous to the profits obtained by banks. ³ Be that as it may, this interpretation is inadequate for grasping the essence of advertising platforms, which do not ‘lend’ data but rather commercialise the audience-commodity produced by the informational labour of a set of workers formally subsumed under these platforms, inaugurating a new and important stage in the development of the Culture Industry (Figueiredo and Bolaño, 2017). According to Dantas and Raulino (2020: 136):

Advertisers do not have access to the databases of SDPs; they only benefit from the mediation in delivering their ads to the chosen audience. They pay for the data but do not privately appropriate it; they do not take it “home” as one might, for example, take home a rug acquired at an antique auction. This economy is similar to the money market, as treated by Marx in Book III, Section IV, of Capital. The bank lends, it does not sell money. The borrower is obligated to return the amount lent, plus interest. In other words, the borrower gains access to a resource necessary for their business but monopolized by another actor – the bank. Therefore, in Marx's terms, the bank's gain takes the form of interest, not strictly profits (gains from the exchange of one commodity for another equivalent commodity, or for money as the universal expression of value). Similarly, in the platform economy, the one who pays for data gains access to it but does not become its owner. The data remains on the platform's servers, available to other “borrowers.” It is a resource of increasing returns, unlike commodities, which are resources of diminishing returns.

Furthermore, Dantas and Raulino (2020) argue that data functions as a source of information that feeds algorithms, analogous to energy sources that sustain a system of machines. According to the authors, ‘data drives the platform, or rather, its algorithms, just as energy drives factory machines or transport ships’ (Dantas and Raulino, 2020: 132). However, this analysis overlooks two important aspects of the problem. First, both data and energy sources require the mediation of labour to be used for other purposes. Second, from the perspective of the labour process, data is highly heterogeneous, while energy sources (electricity, fossil fuels, etc.) are, by nature, homogeneous. As Bolaño (2024: 83) observes:

[…] electricity is a perfectly homogeneous productive input, used in the same way across diverse labour and consumption processes – indeed, a public good with natural monopoly characteristics – something radically different occurs with data, which constitutes a heterogeneous mass whose productive or unproductive use depends on specific needs. For example, Norwegian climate data is useless for defining prison policies in Fortaleza or for developing advertising strategies for the automotive sector in Bangladesh.

In other words, the use of data in specific productive processes depends on the use-value of particular datasets, which are handled by workers with equally specific skills, since what ultimately matters to the capitalist is the use-value of the labour power purchased. Although the authors at one point present a correct summary of Marx's definition of the categories of labour and productive consumption, by adopting the conception that internet users perform unpaid labour, they end up producing a problematic interpretation of these categories. This becomes evident in their comparison with electricity, when they state:

Just as machines need to be connected to some source of energy, algorithms need to be connected to some source of information. For this reason, users must remain in an almost permanent state of activity – an activity which is characterised as productive labour and productive consumption. (Dantas and Raulino, 2020: 132)

The proposals of Fuchs (2014) and of Dantas and Raulino (2020), in suggesting that internet users are working, end up relativising both the exploitation of waged workers and the very concept of social class. These perspectives overlook the fundamental feature of intentionality, intrinsic to all human labour – even in the case of the alienated worker, who loses control over the outcome of their production. Furthermore, treating the consumption of hardware and software by internet users as productive consumption contradicts the Marxian definition of this category, which is restricted to the consumption of the means of labour by the waged worker in the production process. According to Dantas and Raulino (2020: 133):

The activity of PSD users who are at the point of purchase becomes productive labour because it produces the data necessary for the valorisation of capital. Naturally, in this production, these users are also consuming: fixed or mobile terminal devices, telecommunications networks, and the very systems and designs of the platforms. This consumption is productive, just like the consumption of machines and raw materials by workers in any factory. It is consumption that does not ‘annihilate’ the product, although it may wear it down over time (as machines also depreciate), but in the meantime, it produces data valorised in the same way that machines produce textiles.

In sum, Dantas and Raulino's (2020) reference to consumption as annihilation is entirely misplaced, since, according to Marx, all consumption implies annihilation – that is, destruction rather than the creation of value. The properly productive character of so-called productive consumption is not related to whether or not the means of labour are worn down, but to the fact that it coincides with the consumption of labour power in the production process – that is, with the destruction not merely of the use-value and value of any given commodity, but of the worker's very lifetime, for the benefit of capital.

The commodification of crude data and the possibility of fictitious valorisation

The collection and use of data, facilitated by information and communication technologies, can serve specific purposes and be linked to concrete productive processes. In a state-of-the-art industrial plant, such as a car manufacturer, workers using laptops and wearable devices – such as augmented-vision glasses and exoskeletons – generate real-time data (on performance, position, production flow, etc.) that assist in the coordination of labour and in surveillance, directly impacting productivity. An illustrative example is that of Mercedes Benz in São Bernardo do Campo (Pinto, 2020).

We may also cite the case of platforms such as Uber (Zanghelini, 2024; Zanghelini and Bolaño, 2025), which, by taking advantage of the advance of neoliberal policies and the existence of a global mass of unemployed workers, and by owning a specific technical means fuelled by data, exercise extensive external control over the labour process and promote the material subsumption of labour under capital. This enables the platform to parasitically appropriate part of drivers’ income within the sphere of circulation, without establishing a proper wage relation. This specific case, which accentuates the precarisation of labour, constitutes a regressive form of accumulation, similar (though not identical) to the old commercial capital, which externally controlled and coerced artisans in the putting-out system.

Other examples can be found in the production of the audience-commodity by the cultural industries financed through advertising. In this case, it is a specific commodity, the production of which depends on the use of demographic and behavioural data accessed and organised with the support of research companies linked to the advertising market. These data contribute to shaping the audience-commodity, which has a use-value determined for the advertiser and an exchange-value defined by the socially necessary labour time of the cultural workers engaged in its production (Bolaño, 2000). Digital advertising platforms, in turn, are able to produce a much more segmented and personalised audience-commodity than that of the old Culture Industry, exemplified by radio or television (Figueiredo and Bolaño, 2017).

It is also worth highlighting the production of the information-commodity. As already noted, when informational labour acts upon data, these cease to be inert data, merely stored in repositories, and acquire the condition of information – that is, something circulating among humans, with a dynamic and relational character. Information, thus constituted, is not limited to serving as raw material employed in the production of other commodities, as in the case of the audience-commodity; it may also present itself as a commodity in its own right – the information-commodity – endowed with exchange value, from the standpoint of the platforms, and use value for the buyers, which varies according to the context. It may serve, for example, risk assessment, when banks and fintechs employ the information-commodity – produced through the processing of behavioural and consumption data – to set interest rates and credit limits; urban and real estate management, resulting from the processing of georeferenced data that guides decisions on speculation and the planning of new developments; or indeed academic research, which is frequently dependent on large private databases, although it also relies, in many cases, on public databases, in which the information produced by public servants does not take the form of a commodity. In short, the information-commodity is fully integrated into the process of the expanded reproduction of capital.

Beyond all these forms of the utilisation of data in concrete productive processes of goods or services – and linked to them – we advance the hypothesis that the so-called data economy is characterised as an innovative form of fictitious capital, adding to financial derivatives and corporate bonds, whose valorisation rests upon the promise of generating future revenues. In other words, when simply extracted and commercialised by companies that control large repositories, sets of raw data may – though devoid of value – operate as capital, without being capital a priori, thereby fuelling speculation and fictitious valorisation.

This movement constitutes a significant source of income for companies that own digital platforms. The trade of data on specific platforms, such as Data Marketplaces, Data Brokers, and Decentralized Finance (DeFi), as well as Blockchains – though still emerging and poorly structured – constitutes a new form of fictitious capital. Its dynamics follow the same logic as classical forms or other financial innovations that have marked the development of capitalism in the neoliberal period.

Regarding the category of fictitious capital in Marx, Carcanholo and Medeiros (2014) describe it succinctly and accurately in the following terms:

Fictitious capital, in Marx, arises from the dialectical unfolding of what he calls money-dealing capital, which unfolds into interest-bearing capital and culminates in fictitious capital. Historically and logically speaking, fictitious capital is based on the diffusion throughout the economy of the logic of lending capital in exchange for interest (interest-bearing capital). As this logic spreads, the attainment of a certain return, regardless of its source (directly from a loan or a security, no matter the type), is considered as a return derived from capital, whether that capital already exists or not. In other words, under these conditions, capital can literally be constituted from the returns that accrue to its fictitious existence. (Carcanholo and Medeiros, 2014: 293)

The authors further explain:

In summary, fictitious capital is nothing more than speculative relations in which certain speculators force others to engage in chains of debt/credit. In doing so, they create rights of appropriation over values that do not yet exist and may never exist. Today, the reality of this form of capital is so widespread that the market itself recognizes the capitalization that corresponds to it, since the right to appropriation can be resold for a certain capital value, depending on the interest rate. It was following this logic that the process of financial market liberalization took shape from the 1970s onward. The complex process of financial innovations (creation and expansion of financial instruments) is part of this logic, including the development of the famous derivatives market. (Carcanholo and Medeiros, 2014: 294)

Although it is ‘misleading to contrast “speculative” financial markets with a supposedly “solid” capitalist production’, since ‘all capitalist production contains a speculative element’ (Heinrich, 2024: 176), as well as reducing the problem to regulating financial capital to harmonise it with industrial capital (Grespan, 2015), with the advance of neoliberal globalisation there has been a growing imbalance between the volume of fictitious capital in circulation and what is effectively materialised in the sphere of value production. The 2008 crisis itself, incidentally, ‘is nothing more than a limit situation of this tragedy and is explained precisely by the predominance of the dysfunctionality of the logic of fictitious capital for the accumulation of total capital’ (Carcanholo and Medeiros, 2014: 295).

Thus, even though fictitious capital is configured as a dialectical process, in which functionalities and dysfunctions coexist, it can be observed that, when the expected flow of surplus value is not generated in the productive sphere, ‘an increasingly larger portion of global capital will seek to appropriate a value that is being produced less and less. The final result is the reduction of the profit rate and the deepening of the cyclical behaviour of the crisis’ (Paschoal and Carcanholo, 2009: 10).

Therefore, by promoting the securitisation of a potential capital-value that exceeds the process of real accumulation, fictitious capital plays no role in containing the law of the tendential fall in the rate of profit; on the contrary, it tends to intensify the overproduction of capital which, in a context of declining profit rates, finds ever less room for valorisation. For capital to regain valorisation in a new cycle of development and accumulation, its devaluation becomes necessary, which gives the crisis even more dramatic contours. Indeed, the fall in the profit rate is not the cause of the crisis, as some interpretations in the literature suggest, but rather the manifestation of a crisis in which capital-value has been produced in excess (Carcanholo, 2011). In summary:

This growing mass of fictitious capital reinforces the dialectical character of the capital accumulation process. On the one hand, it creates a space for the valorisation of over-accumulated capital that, without this alternative, would provoke a reduction in profit rates. On the other hand, due to the very nature of fictitious capital, it is capital that does not directly produce (additional) value. Therefore, its expansion means the proliferation of titles of appropriation over a value that is not necessarily produced pari passu. The consequence, in this regard, is the reduction of profit rates, a clear manifestation of the crisis. (Carcanholo and Medeiros, 2014: 294–295)

In the next section, we move to a different level of analysis. At this stage, the aim is no longer to examine the specific circumstances in which data may or may not be linked to the productive process and the generation of value, nor to point to the possibility of its fictitious valorisation. The focus shifts to the national question, taking as a reference the use value of data and the ways in which its control can strengthen major oligopolies beyond commodity production and speculation – serving, above all, as a geopolitical and strategic instrument in the struggle for power and global hegemony. This becomes particularly relevant in the current scenario, in which ‘geopolitical competition revolves centrally around the control of information and communication technologies’, more specifically ‘the tensions between the United States and China over the so-called fifth-generation internet (5G) and, more recently, over what is now referred to as artificial intelligence (AI)’ (Martins and Lopes, 2024: 27).

The use-value of data and the question of national sovereignty

D’Alva and Paraná (2024) present an important discussion on the current trend, supported by the UN's 2030 Agenda, of using big data by national statistical institutions through public–private partnerships with large digital platforms and other companies. Without delving into the details of this valuable contribution, we highlight only the fact that the capacity of the national state to control its data is at stake. The authors observed, in the empirical cases studied, not only a tendency for national institutions – centuries-old and highly qualified – to lose control over data but also signs of resistance from the technical staff interviewed regarding alliances that result in the transfer of digitised information related to official statistics produced by the state to technology companies. This scenario raises concerns about the ability to formulate national development policies with the necessary autonomy emphasised by Furtado (1978). According to D’Alva and Paraná (2024: 9)

In its etymology, statistics means ‘science of the state’ and its historical origins relate back to the processes of the constitution of the states themselves. The birth of the territorial states is inseparable from an immense accumulation of informational capital, provided mainly by statistics. From the monarch's administration secrets, the official statistics of the nation-states became a public good controlled by a statistical field. This process followed the qualitative transformation of a series of private capital into public capitals that characterised the relative autonomisation of modern bureaucracy. The nation-state has ever since been associated with a ‘total’ knowledge of the social world, which is provided precisely by statistics. This status and its associated symbolic power have been challenged in the present by new forms of knowledge and action controlled by transnational tech corporations.

The case of official statistics, though emblematic and crucial, represents only one part of the broader problem we are currently facing, as digital networks and platforms have taken on a central role in the mode of regulation of capitalism, deepening trends established since the beginning of the neoliberal period, in the wake of the extension of the logic of value into the most remote areas of human relations. In the context where large multinational corporations store personal and institutional data, Sergio Amadeu da Silveira (2024) points out, when discussing so-called cloud computing, how the 2024 cyber blackout of Microsoft's operating system is evidence of the power – and potential harm – exercised by large digital platforms:

Amazon Web Services and Microsoft Azure, in 2021, held 60% of the global cloud market, offering infrastructure as a service. What does this mean? That various companies, institutions, and governments have replaced their own local data processing and storage infrastructures with contracts for Amazon and Microsoft to “manage” and “rent” data storage space and computational services […] The blackout demonstrated the enormous power held by a mediator of digital relations and a data processing operator like Microsoft. Without a doubt, the unintentional failure caused the blackout. But it is evident that Microsoft has the power to block access by companies and institutions to their own data located in its data centers, far from our jurisdiction and our capacity for physical access.

Marcio Pochmann (2024a) argues that the international division of labour has been renewed in the digital age, intensifying the process of unequal exchanges between countries. This would inaugurate a new stage of capitalist underdevelopment, as unequal exchanges would now occur between ‘data-extracting and processing countries’ and those that are limited to being ‘suppliers of raw data’ (Pochman, 2024a: 40). According to the author, in the old industrial era, when the ten largest U.S. companies were firms like Esso, Ford, and General Motors, ‘the backwardness of a country was noted by the absence of industrial companies, with the unequal exchange relationship located in the export agenda based on primary commodities’ (Pochman, 2024a: 40). Now,

[…] raw data has taken the form of a commodity. This is because, in the first third of the 21st century, datafication has been the means through which the governance of populations and territories that prevails in the world has become the object of dispute in a public sphere increasingly dominated by private interests. (Pochman, 2024a: 40)

Although the problem of unequal exchange is intensified by the growing complexity introduced by the digitalisation paradigm – since its essential dynamic is linked to differing organic compositions of capital, generating disparities in the value produced and appropriated between technologically advanced countries and those in subordinate positions (Carcanholo and Amaral, 2008; Dussel, 1988; Marini, 2000) – we believe Pochmann is mistaken in suggesting that this process can be explained by treating data as primary commodities. It is true that peripheral specialisation in the production of primary commodities reflects, under the law of the tendential equalisation of the average rate of profit, a transfer of value to the centre. However, this reasoning cannot be directly applied to the appropriation of data from the periphery by firms based in the centre, since data do not possess intrinsic value.

Thus, the digital traces left by internet users in peripheral countries do not constitute a direct and automatic source of revenue for the major technology companies based in the United States. The problem identified by Pochmann (2024a) – namely, that underdeveloped countries act as ‘mere exporters of raw data’, forming a ‘new source of income operated by large private tech companies’, whose revenues ‘exceed the GDP of several countries’ and remain ‘virtually unregulated and untaxed’ (p. 42–43) – is, indeed, real. However, to fully grasp the underlying logic, it is necessary to understand the mediations and the concrete forms that enable not merely the transfer of the data themselves – crude material devoid of value – but rather the transfer of value between the centre and the periphery.

A first response, in this regard, stems from the fact that the major oligopolistic companies that own digital platforms – predominantly US-based – expand their speculative capacity, as discussed in the previous section, by appropriating data from a significant portion of the planet and thereby increasing the volume of fictitious capital at their disposal. Secondly, these data can be used in the production of commodities, particularly within the cultural industry. However, it is necessary to examine, in each case, the specific centre – periphery relations involved – although, in general, it is this same global big tech oligopoly that stands to reap the benefits of any outcomes obtained. Thirdly, all of this reinforces the economic and geopolitical power of these actors, with consequences for the processes of uneven development among nations and for threats to sovereignty that extend beyond mercantile exchange relations.

Morozov (2023), for his part, highlights that the strategy of companies based in Silicon Valley – with which the Brazilian state will certainly have to deal in its effort to confront dependency on digital technologies – consists of starting their operations in a single specific area and then diversifying into many others. As the author observes, there is a significant strategy on the part of U.S. technology companies ‘to enter the fields of health, education, and national security. They started as content distribution mechanisms, merely organising information and selling advertising. Now, they have become a gateway to almost everything’ (Morozov, 2023). This reflects, we may add – drawing on van Dijck's (2022) tree metaphor for the platform ecosystem – that the so-called big techs, represented by the trunk, are expanding into a wide range of branches, that is, into the most diverse sectors of the economy, both public and private.

To confront this strategy of large technology corporations, Morozov (2023) suggests two paths of action. The first, less effective, consists of the state ‘imposing restrictions on the data that can be used for, for example, generative artificial intelligence’. To some extent, Brazil's General Data Protection Law (LGPD), approved in 2018, represents a step in this direction, establishing rules for the processing of personal data. The second, more combative, proposes that the state develop ‘robust public infrastructure that can encompass as many layers of these digital systems as possible’. We can consider that it is in this sense that Pochman (2024a) advocates for the creation of the National System of Geosciences, Statistics and Data:

The Revolution of 1930 grounded economic sovereignty in the strategy of import substitution for manufactures. Almost a hundred years later, data sovereignty is at stake amid the country's potential to establish its own data infrastructure through a sovereign cloud capable of integrating existing databases and, above all, new ones generated by the digital era itself. The creation of the National System of Geosciences, Statistics, and Data would define a horizon for the Brazilian government's strategy to reaffirm data sovereignty. In doing so, it would break with the condition of being a producer of primary data commodities, confronting yet another form of underdevelopment. (Pochman, 2024a: 43)

In the case of the public–private partnerships studied by D’Alva and Paraná (2024), the problem is evident. The authors, both in the cases they studied and in other examples from the literature, highlight the difficulties that national statistical offices (NSOs) face in accessing private data, indicating that this is precisely ‘evidence of the contemporary process of data enclosure, which constitutes an element of a broader dispute between nation-states and private corporations over control of informational capital’ (D’Alva and Paraná, 2024: 11). By using the concept of enclosure, the authors follow a line similar to that of Bolaño (2004), who applies it to genetic mapping, diverging from the perspective of Dantas and Raulino (2020), previously cited, whose analysis associates enclosures with the patent system, with reference to algorithms. Be that as it may, under no circumstances can data themselves be patented – only knowledge.

In general terms, D’Alva and Paraná (2024) define the concept of ‘data enclosure’ as the process in which ‘information about user activity generated by digital interactions is kept away from the users themselves and the public for the benefit of service providers (platforms and technology companies) where the data is generated’

By this, open or shared arrangements of access and control over data are made proprietary and exclusive. Enclosed, big data constitutes informational capital under dispute and a digital commodity in potential. Therefore, while those in the statistical field seek access to data to maintain their positions as relevant producers of statistical information, for the private sector the data enclosures are necessary for enabling new data markets. (D’Alva and Paraná, 2024: 11)

It is curious to note that the reference to the enclosure process, often evoked in debates on copyright and related topics, was common at the turn of the twenty-first century, referring to a kind of enclosure promoted by the international human genome mapping project. In this context, an epic battle was fought between the original public perspective and that of the company Celera Genomics (Bolaño, 2024). What is interesting is that it involved the global cooperative construction of large databases – genetic and informational – that would form the basis of genomic studies in the following decades (Bolaño, 2004). In all cases, the disjunction between public and private interests opens up in defining the approach to be adopted in implementing technical innovations with significant impact, as is the case with information and communication technologies, responsible for the constitution of large databases and algorithm systems, both of great interest for planning and social policies.

Perhaps the decisive difference between the private Human Genome Project and the major digital oligopolies lies in the fact that, in the former, what is at stake is the control of knowledge with real potential for the development of the health industry; whereas in the latter, it is the control of vast volumes of digital data that perform multiple functions, both in the process of capital accumulation and in the management of public policies and in the assertion of the political-strategic power of nation states.

Final considerations

Initially, we explain that the generalisation of data collection and use derives from the advance of the productive forces in the context of the Third Industrial Revolution, whose developments have profoundly reconfigured labour processes. Furthermore, we argue that data are crude raw material without value until mobilised by informational labour, and we criticise interpretations that attribute to internet users the condition of unpaid productive workers. We then advance the hypothesis that the speculative commodification of raw data, in certain markets, may take shape as new forms of fictitious valorisation of capital. Finally, we examine the geopolitical dimension of the digital sphere, highlighting the concentration of data in large platforms, the so-called Big Techs. The exploitation of data extracted from the population by private companies constitutes, in fact, a form of expropriation of a public good and, when carried out by large foreign corporations operating in the productive, commercial or speculative spheres, represents a concrete risk to the security and sovereignty of peripheral countries.

Complementary Bill No. 234/2023, for example, which is under consideration in the Brazilian Chamber of Deputies and known as the General Law of Data Empowerment, aims to create a Brazilian Ecosystem of Data Monetisation, allowing individuals to trade their personal data in the market (Cunha et al., 2025). The proposal draws inspiration from the regulation of Open Finance, ⁴ which authorises financial institutions to share customer data with consent (Cunha et al., 2025). In other words, the bill provides for the centralisation of data in private companies, subject to taxation, responsible for commercialising them and remunerating their holders. Such an arrangement would establish ‘a new “paradigm” in the relationship between data holders, technology companies and public authorities’ (Cunha et al., 2025), with commercialisation taking place, as we have discussed throughout the text, either in the form of the sale of the information-commodity or speculatively, the latter being the outcome most likely to be favoured by the creation of such a market.

A practical rehearsal of this public–private logic, albeit without concrete implementation, took place in April 2025, when Dataprev (the Social Security Technology and Information Company) announced a partnership with DrumWave to transform data – initially those from payroll loans – into ‘economic assets’. The proposal envisaged that data would not only circulate but also be recorded and stored in a digital wallet, enabling subsequent transactions and advantages for their holders (Cunha et al., 2025). This example clearly illustrates the point we have defended: data in themselves have no value, although they may acquire a price – as in the case of virgin land in Marx.

In sum, the State institutions responsible for producing national statistics have as their primary function the creation of databases necessary for planning, with a view to the country's development and the wellbeing of the population. Any attempt to commercialise such data, in our view, places national sovereignty at risk, with unpredictable consequences. Indeed, the very management of data by the State is only acceptable insofar as citizens trust the technical bodies in charge, which must guarantee both statistical confidentiality and ‘de-named’ access to data – a ‘basic rule for working with information from a statistical point of view’ (Pochman, 2024b), aimed at improving public management in the service of citizenship and good living.